一种用于自适应噪声抵消的变步长LMS算法

为了提高LMS自适应滤波算法的性能，在分析已有变步长算法的基础上进行了一些改 进。改进算法用误差信号的自相关来调节步长以实现对不相关噪声的更好抑制，且采 用先固定后变化的方法控制步长，兼顾了暂态和稳态性能。利用改进算法进行了自适应噪声 抵消的仿真实验,结果表明，基于改进变步长LMS算法的自适应噪声抵消器 能有效抵制噪声干扰，对含噪信号具有良好的消噪能力。

A Variable-Step LMS Algorithm for Adaptive Noise Canceller

To improve the performance of LMS adaptive filtering al gorithm, an improved variable-step LMS algorithm is proposed based on analysis of existin g algorithms.With the autocorrelation step regulation of error signal, the impro ved algorithm realizes the better uncorrelated noise inhibiting ability. It gives c o nsideration to both transient and steady performance by the step control method of fi rst fixation and last variation. The simulation experiments of adaptive noise ca ncellation illustrate that the adaptive noise canceller based on modified variab le step LMS algorithm has a better de-noising performance for the noise signal.